Recent molecular phylogenetic work has found that Breviceps Merrem, 1820 comprises two major clades, one of which, the B. mossambicus group, is widely distributed across southern sub-Saharan Africa. This group is notable for harboring abundant cryptic diversity. Of the four most recently described Breviceps species, three are members of this group, and at least five additional lineages await formal description. Although Breviceps has long been known to occur in Angola, no contemporary material has been collected until recently. The three most widespread taxa, B. adspersus, B. mossambicus, and B. poweri, may all occur in Angola, but accurate species assignment remains challenging given the rampant morphological similarity between these taxa, and, until recently, the lack of genetic resources. Phylogenetic, morphological, and acoustic analyses of recently collected samples from disparate localities within Angola provide evidence for an undescribed species that is sister to B. poweri. The new species can be diagnosed from its sister taxon by lacking pale spots along the flanks, a pale patch above the vent, and a short, dark band below the nares (all present in B. poweri). Additionally, the male advertisement call differs from the three other Breviceps that might occur in Angola in having both a longer interval between consecutive calls and a higher average dominant frequency. We here describe this lineage as a distinct species, currently only known from Angola, and discuss the presence of other Breviceps taxa within Angola.

Investigações moleculares recentes revelaram que o género Breviceps Merrem, 1820, é composto por duas linhagens principais, uma das quais, o grupo B. mossambicus, é amplamente distribuído na região sul da África subsaariana. Este grupo é notável por albergar uma abundante diversidade críptica. Das quatro espécies de Breviceps recentemente descritas, três pertencem a este grupo, e pelo menos outras cinco linhagens adicionais aguardam a sua descrição formal. Apesar de o género ser conhecido de Angola desde há muito tempo, só muito recentemente foram colhidos novos espécimes. Os três taxa mais amplamente distribuídos, B. adspersus, B. mossambicus e B. poweri podem todos, porventura, ocorrer em Angola, no entanto a correta identificação destas espécies têm sido problemática devido às semelhanças morfológicas extremas entre este taxa, e, até muito recente, a completa ausência de material genético. Análises filogenéticas, morfológicas e acústicas dos espécimes recentemente colhidos em diferentes locais de Angola apontam para a existência de uma espécie nova para a ciência, irmã de B. poweri. A nova espécie pode ser diferenciada do seu táxon irmão pela falta de marcas pálidas nos flancos, mancha pálida acima do ventre e pequena banda negra abaixo do nariz (presentes em B. poweri). Para além destas características, o chamamento dos machos difere das outras três espécies de Breviceps que podem ocorrer em Angola por ter um maior intervalo entre chamamentos consecutivos e uma maior frequência média dominante. Descrevemos aqui esta linhagem como uma espécie distinta, atualmente apenas conhecida de Angola, e discutimos a presença de outras espécies de Breviceps em Angola.

Afrobatrachia, Anura, Breviceps ombelanonga sp. nov., cryptic species, multilocus, novel species, Sub-Saharan Africa

África Subsahariana, Afrobatrachia, Anura, Breviceps ombelanonga sp. nov., espécies crípticas, espécies novas, multilocus

Breviceps Merrem, 1820 is a genus of fossorial frogs widely distributed across southern sub-Saharan Africa, from Angola in the northwest, through Zambia, the southern portions of the Democratic Republic of the Congo and Tanzania, and southward throughout virtually all of southern Africa (Minter 2004; Minter et al. 2017). It currently comprises 18 species, although a recent molecular phylogenetic study indicates that this is an underestimate (Nielsen et al. 2018). Six species have been described since 2003 (Minter 2003; Channing and Minter 2004; Channing 2012; Minter et al. 2017), largely representing cryptic taxa embedded within what were previously considered widespread species or species complexes, namely B. mossambicus Peters, 1854 and B. adspersus Peters, 1882 (Nielsen et al. 2018). The justification for recent descriptions has largely been variation in nuptial call characteristics, geography, and mitochondrial genetic distances, yet many additional distinct genetic lineages have been identified and await formal description. Large-scale taxonomic revision is required but this remains problematic due in large part to limited genetic sampling (Nielsen et al. 2018), especially in the northwestern extent of the genus in Angola.

The taxonomy of Angolan Breviceps has long been problematic. Bocage (1870, 1873) was the first to report Breviceps in Angola based on two specimens from “Biballa” (currently Bibala, Namibe Province) that he referred to Breviceps gibbosus (Linnaeus, 1758). After receiving more specimens from other localities in Huambo and Huíla provinces, Bocage (1895) provided a more detailed description of the Angolan material and assigned all of these records to B. mossambicus. He noted that compared with other Breviceps (which, at the time, included only three species), Angolan specimens lacked a heavily granular dorsum (vs. granular in B. verrucosus) and had a continuous dark gular patch (vs. paired patches in B. adspersus). Unfortunately, the majority of these specimens were lost in the 1978 fire that destroyed the Lisbon Museum (Almaça 2000; Marques et al. 2018). Subsequent workers provided additional records from western Angola (Bengo Province: Parker 1934; Huambo and Huíla provinces: Monard 1938; Benguela Province: Monard 1938, Helmich 1957) and northeastern Angola (Lunda-Sul and Moxico provinces: Laurent 1964; Ruas 1996), all of which were reported as B. mossambicus. In a second review of the same material, Ruas (2002) revised her previous conclusions and referred the specimens from Moxico Province to the “Breviceps mossambicus-adspersus complex” (sensu Poynton 1982; Poynton and Broadley 1985), noting genetic data were needed to resolve their taxonomy. This species complex has been suggested to have a broad hybridization zone across southern Africa (Poynton 1982), and Angolan Breviceps were noted to share aspects of coloration with both B. mossambicus and B. adspersus, yet were distinct from B. poweri Parker, 1934 from the Zambezi Basin (Poynton and Broadley 1985). More recent synopses of Angolan material have either referred historical material to B. cf. adspersus (Baptista et al. 2019) or simply as B. sp. in recognition of the taxonomic uncertainties for these populations (Marques et al. 2018; Ceríaco et al. 2020).

A recent phylogenetic study of Breviceps (Nielsen et al. 2018), while lacking Angolan material, confirmed the presence of B. poweri in northwestern Zambia, as well as nomintotypical B. adspersus within 3 km of the Angolan border in Namibia (Fig. 1A). This suggests that both might also occur in Angola (Marques et al. 2018), although the evidence for B. poweri is based mainly on tertiary references (see Channing and Rödel 2019). Based solely on external morphology, Ceríaco and Marques (2018) recently identified specimens from Moxico Province, in eastern Angola, as B. poweri; these are the same specimens previously identified by Ruas (1996, 2002) as B. mossambicus and B. mossambicus-adspersus, respectively. While B. mossambicus has been historically listed as part of the Angolan anuran fauna, recent genetic analyses have so far only confirmed populations from Mozambique as corresponding to this name (Nielsen et al. 2018). Due to substantial morphological similarity, scarcity of genetic sampling, and potential for hybridization among B. mossambicus, B. poweri, and B. adspersus (Poynton 1964, 1982; Poynton and Broadley 1985; Minter et al. 2017), taxonomic identification of any historical Angolan material should therefore be considered tentative at best.

Angola’s long civil war, which lasted from 1975 to 2002, effectively stifled biological exploration and discovery (for additional summary, see Marques et al. 2018). Recent surveys, many by authors of this manuscript, have produced the only contemporary records of Angola’s herpetofauna (e.g., Ceríaco et al. 2014, 2016, 2018; Conradie et al. 2016; Heinicke et al. 2017; Marques et al. 2018; Baptista et al. 2019; Butler et al. 2019; Ernst et al. 2020), including the only recent records of Breviceps in Angola. The nearest samples with confident identifications and associated genetic data are at least 600 km away (i.e., B. adspersus in Namibia and B. poweri in Zambia; Nielsen et al. 2018). Here we analyze these recently collected Angolan Breviceps in a phylogenetic framework and assess their taxonomic status, resulting in the description of a new species so far known only from Angola.

Our phylogenetic analyses indicate that sampled individuals from Angola form a clade that is genealogically exclusive from other described species of Breviceps (Fig. 1). These populations are morphologically diagnosable from other closely related taxa, specifically possessing distinct coloration and pattern that differ from the sister taxon, B. poweri. A PCA of mensural and meristic data indicates that the Angolan individuals fall within a unique region of morphospace, with a relatively narrower head, shorter thigh, and shorter manual digit III in comparison to closely related species. Lastly, there are distinct acoustic differences associated with the male nuptial call. Thus, we here describe these populations as a new species.

Breviceps ombelanonga sp. nov.

http://zoobank.org/E3815018-4176-4073-92B8-E65274D354FB

Figs 4, 5, 6, 7 Suggested common names: Angolan Rain Frog (English), Sapinho das Chuvas de Angola (Português).

Chresonymy.¹

Breviceps gibbosus: Bocage (1870: 68).

Breviceps gibbosus: Bocage (1873: 227).

Breviceps mossambicus: Bocage (1895: 182); Parker (1934: 194); Monard (1937: 29, 1938: 56); Laurent (1964: 156); Cei (1977: 17, 18); Ruas (1996: 23).

Rana mossambicus: Hellmich (1957: 30).

Breviceps “mossambicus-adspersus” complex: Poynton (1982: 67); Ruas (2002: 142).

Breviceps adspersus [part]: Poynton and Broadley (1985: 52).

Breviceps sp.: Marques et al. (2018: 81); Ceríaco et al. (2020: 63).

Breviceps cf. adspersus: Baptista et al. (2019: 270).

Material examined

Holotype. UF Herp 187172 (field number MCZ A-36476), an adult male, Kawa Camp Headquarters, 1 km south of the Kwanza River, Kissama National Park (-9.183068, 13.369314, WGS-84, elevation 151 m above sea level), Luanda Province, Angola, collected by LMPC, Mariana P. Marques, Philip Pastor, and John Cavagnaro on 2 June 2016 at approx. 22:00. Paratypes (5 males, 1 female, 1 sex unknown) UF Herp 187173 (field number MCZ A-36495), an adult male, Kawa Camp Headquarters, 1 km south of the Kwanza River, Kissama National Park (-9.183068, 13.369314, WGS-84, elevation 151 m above sea level), Luanda Province, Angola, collected by LMPC, Mariana P. Marques, Philip Pastor, and John Cavagnaro on 8 June 2016; MHNCUP/ANF 320 (field number AMB 11736), sub-adult (sex unknown), Embala Seque, 14 km N of Cassumbi village (-11.083845, 16.66741), Bié Province, Angola, collected by LMPC, Mariana P. Marques, and Adam Ferguson on 16 June 2019; PEM A12800 (field number WC-4591), adult male, Cuanavale River source lake (-13.089343, 18.89485, 1396 m above sea level), Moxico Province, Angola, collected by Werner Conradie and Luke Verburgt on 24 October 2016; PEM A12537 (field number WC-3924), juvenile male, Cuito River source lake (-12.68935, 18.36012, 1435 m above sea level), Moxico Province, Angola, collected by Werner Conradie and Ninda Baptista on 18 February 2016 October; PEM A12787 (field number WC-4756), adult male, Quembo River source lake (-13.13544, 19.04397, 1375 m above sea level), Moxico Province, Angola collected by Werner Conradie on 11 November 2016; PEM A12770 (field number WC-4827), adult female, Cuando River source (-13.00334, 19.13564, 1364 m above sea level), Moxico Province, Angola, collected by Werner Conradie and James Harvey on 22 November 2016; SAIAB 204537 (field number Ang16-RB12), adult male, Quembo River source lake (-13.13583, 19.04528), Moxico Province, Angola, collected by Roger Bills on 9 November 2016.

Diagnosis

A species referable to Breviceps due to the following characteristics (Poynton 1964; Minter et al. 2017): snout extremely abbreviated; mouth narrow and downturned near jaw joint; short limbs which, at rest in life, are held close to the body, not projecting beyond the body outline; digits I and V short or rudimentary; inner metatarsal tubercle well developed and notably longer than pedal digit III, narrowly separated from a prominent conical outer metatarsal tubercle. Additionally, the results of the molecular phylogenetic analyses support this species as embedded within the diversity of Breviceps, specifically within the B. mossambicus group (Fig. 1B). Breviceps ombelanonga can be diagnosed from other species of Breviceps and especially those in the B. mossambicus group by the combination of lacking a visible tympanum, males having a single, uniformly dark gular patch that is continuous with the mask extending from the eye, having generally smooth dorsal skin, lacking many small tubercles on the palmar surfaces (as in, e.g., B. branchi and B. sylvestris; FitzSimons 1930; Channing 2012), lacking pale spots along flanks and a pale patch above the vent (both present in B. poweri; Parker 1934; du Preez and Carruthers 2017), lacking short dark band below nares (as in B. poweri; du Preez and Carruthers 2017), lacking confluent inner and outer metatarsal tubercles, having a relatively narrower head, shorter thigh, and shorter manual digit III (Fig. 2; Table 4), and having an advertisement call with both a longer interval between consecutive calls and a higher average dominant frequency (Fig. 3).

The new species can be distinguished from other species of Breviceps occurring in the region by the following: pale paravertebral and dorsolateral patches are lacking, although a fine dorsolateral band may be present (versus no pale paravertebral or dorsolateral spots or patches in B. mossambicus; series of both paravertebral and dorsolateral pale spots and patches present in B. adspersus, a series of pale dorsolateral spots or patches present in B. poweri); no conspicuous light patch above vent (present in B. poweri); manual digit IV reaching approximately midway between the proximal and distal subarticular tubercles of manual digit III (versus nearly reaching distal subarticular tubercle of manual digit III in B. mossambicus; not reaching or barely passing the proximal subarticular tubercle of the manual digit III in B. poweri; similar to B. adspersus in usually not reaching the distal subarticular of manual digit III); gular region with a single uniformly dark patch (versus a pair of marbled to freckled patches in B. adspersus).

The advertisement call of the new species (Table 5, Fig. 3) can be differentiated from other potential Angolan congeners by its duration (0.175 ± 0.083 s; shorter than in B. adspersus 0.196 ± 0.047 s and B. mossambicus, 0.500 ± 0.070 s, and longer than in B. poweri, 0.140 ± 0.012 s), longer interval between consecutive calls (0.996 ± 0.133 s; B. adspersus, 0.745 ± 0.636 s; B. mossambicus, 0.710 ± 0.168 s; B. poweri, 0.743 ± 0.166 s), and a higher dominant frequency (2156 Hz; B. adspersus, 1742 ± 100 Hz; B. mossambicus, 1835 ± 107 Hz; B. poweri, 1728 ± 83 Hz). The number of pulses per call (28–34) are similar to B. adspersus (14–31), B. mossambicus (7–31), and B. poweri (10–74).

Table 5.

Download as

CSV

XLSX

Comparison of the main variables for the advertisement calls of Breviceps ombelanonga sp. nov., Breviceps mossambicus, Breviceps adspersus and Breviceps poweri. Comparative data taken from Minter (1997, 2003).

	B. ombelanonga sp. nov.		B. adspersus		B. mossambicus		B. poweri
	avg ± sd	range	avg ± sd	range	avg ± sd	range	avg ± sd	range
Call duration (s)	0.175 ± 0.083	0.064–0.342	0.196 ± 0.047	0.077–0.293	0.500 ± 0.070	0.036–0.079	0.140 ± 0.012	0.111–0.160
Call interval (s)	0.996 ± 0.133	0.742–1.190	0.745 ± 0.636	0.363–0.745*	0.710 ± 0.168	0.396–1.17	0.743 ± 0.166	0.500–1.100
No. of pulses/call	30 ± 2.6	28–34	23 ± 3.3	14–31	9 ± 1.2	7–13	30 ± 16.3	10–74
Dominant frequency (Hz)	2156	na	1742 ± 100	1482–2179	1835 ± 107	1600–2193	1728 ± 83	1557–1903

Description of the holotype

Adult male (SUL 30.5 mm), with globular body and well-developed short limbs with medialmost and lateralmost digits reduced (Fig. 4; Table 3); snout abbreviated, protruding and angular in lateral profile, blunt and rectangular in dorsal view; eyes projecting beyond profile of head in both dorsal and ventral views; pupils horizontally elliptical; nares small oval slits, directed horizontally and visible in dorsal and lateral views; mouth narrow and directed ventrally near jaw joint; choana largely obscured by maxillae in ventral view; well-developed gland at midline of palate between choana; tongue ovoid and filling floor of mouth, and lacking median papilla; single medial bony point on lower jaw at symphysis; tympana not distinguishable; teeth absent on premaxilla, maxilla, and vomer.

Figure 4. 

Breviceps ombelanonga sp. nov. holotype male (UF Herp 187172): A in life photo B dorsal and ventral aspects C additional views of the holotype, including the left pes, frontal, right lateral, and left manus and mental. Scale bar: 10 mm. Photographs by J. Cavagnaro (A) and SVN (B, C).

Skin of dorsum and head smooth, and weakly glandular with irregular folds; skin of ventrum smooth; skin folds overlying vent creating triangular shape.

Limbs short with digits I and V short or rudimentary; webbing absent on manus and pes; nuptial pads absent and adhesive glands not discernable; relative manual digit lengths when adpressed: III>II>I>IV; only tip of first pedal digit extending beyond fleshy webbing and sole; fourth (outer) manual digit reaches midway between the large tubercle at metacarpophalangeal joint and subarticular tubercle at most proximal interphalangeal joint; finger tips conical, not expanded; several small globular palmar tubercles; single subarticular tubercles present on pedal digits II, III, and IV; pedal digit V very short, falling short of most proximal subarticular tubercle of pedal digit IV; well-developed (though not keratinized) inner metatarsal tubercle visibly longer than pedal digit III, separated from conical outer metatarsal tubercle by deep cleft.

Coloration

In life, dorsum of body mottled dark brown on pale tan base, transitioning to golden yellow on the lateral aspects, before stark transition to solid dark brown flanks with a dark boundary becoming paler ventrally (Fig. 4); limbs dark grayish brown dorsally; plantar and palmar surfaces pale grayish brown; subarticular, palmar, and inner and outer metatarsal tubercles pale gray; posterior dorsum dark gray-brown with scattered pale gray spots; bold facial mask composed of broad dark brown stripe running obliquely downwards, from margin of lower eyelid towards base of arm (but not attaining it) and joining dorsolateral aspect of gular patch, giving appearance of a large dark bib; region below nares generally same coloration as dorsal and lateral rostrum, and not more darkly pigmented; lower eyelid with white opaque patch at anterior margin; margins of mouth and lateral angle of mouth off-white to cream; gular patch uniformly dark anteriorly, becoming mottled posteriorly and merging with ventral coloration; pectoral region and ventrum creamy pale gray with scattered punctate gray dots sometimes coalescing into larger spots in the gular region and laterally; iris bright orange, scattered with dark brown flecks (dark brown in preservative), with black pupil (pale gray in preservative; no mid-vertebral line; faint pale line extending across posterior hindlimbs extending between heels.

In preservative, coloration is largely similar but more muted and overall darker (Fig. 4).

Measurements

Measurements of the type series are shown in Table 3.

Variations

All specimens resemble the holotype in the absence of a visible tympanum, and skin that is densely granular dorsally and laterally and smooth ventrally (Figs 5–6). The distal tip of manual digit IV reaches well past the proximal subarticular tubercle of manual digit III in all specimens. PEM A12770 have both manual digit II and III proximal subarticular tubercles divided. Inner and outer metatarsal tubercles not separated by a deep cleft in paratypes PEM A12800, PEM A12537, PEM A12787, PEM A12770 and SAIAB 204537.

Figure 5. 

Variation in color and pattern within living paratypes of B. ombelanonga sp. nov.: A, B sub-adult (of unknown sex) from Embala Seque (14 km N of Cassumbi village), Bié Province (MHNCUP_ANF 0320) C juvenile male, Cuito River source lake, Moxico Province (PEM A12537) D adult female, Cuando River source, Moxico Province (PEM A12770) E adult male, Quembo River source lake, Moxico Province (PEM A12787) F adult male, Cuanavale River source lake, Moxico Province (PEM A12800). Photographs by LMPC (A, B) and WC (C–F).

Color and pattern in UF Herp 187173 is very similar to the holotype. Dorsum gray with scattered black spots (MHNCUP/ANF 320); red with scattered black blotches in two specimens (PEM A12537 and PEM A12770), dark brown to black with red spots and markings (PEM A12787 and PEM A12800), light brown with red spots and darker black blotches (SAIAB 204537). Interocular bar visible in all paratypes, except PEM A12537, PEM A12770 and SAIAB 204537. Light dorsolateral patches present in PEM A12878, absent in PEM A12770, dark black band present in PEM A12537. Mid-vertebral line present in most paratypes, but very faint in PEM A12537 and PEM A12770, and absent in SAIAB 204537 and MHNCUP/ANF 320. Heel-to-heel line present in all specimens, but faint in UF Herp 187173, PEM A12537, and PEM A12770. A broad, black stripe runs obliquely downwards from margin of lower eyelid towards base of arm, not reaching the shoulder in all specimens. Dark orbital band partly reaching the gular patch in all specimens, falling short in PEM A12770 (female). Anterior to the orbital bar, a broad white stripe runs down to angle of mouth and onto upper and lower lips in all individuals. Gular patch uniform dark brown to black in all paratypes, except PEM A12537 in which it is dark brown with scattered darker blotches. Pectoral region white, with scattered spots in all specimens. Ventrum white with scattered darker spots in all paratypes.

Figure 6. 

Variation in color and pattern within preserved paratypes of B. ombelanonga sp. nov.: A an adult male from Kawa Camp Headquarters, Luanda Province (UF Herp 187173) B sub-adult (of unknown sex) from Embala Seque, Bié Province (MHNCUPANF 320) C adult male from Cuanavale River source lake, Moxico Province (PEM A12800) D juvenile male from Cuito River source lake, Moxico Province (PEM A12537) E adult male from Quembo River source lake, Moxico Province (PEM A12787) F adult female from Cuando River source, Moxico Province (PEM A12770); and G adult male from Quembo River source lake, Moxico Province (SAIAB 204537). Photographs by SVN (A), LMPC (B), and WC (C–G).

Advertisement call

The following call description is based on a recording of a paratype male (PEM A12800) from the source lake of the Cuanavale River recorded on 24 October 2016 at 8:50 in the morning. Ambient temperature was not recorded. Frogs began calling during the daytime following heavy rains, and stopped after sunset. Call sites were among leaf litter in dense miombo woodland. The call can be described as a short whistle with a call duration of 0.064–0.342 seconds and call interval of 0.742–1.190 seconds. Each call consists of about 28–34 pulses and a dominant frequency of 2156 Hz (Table 4, Fig. 2). The small number of calls from a geographically restricted sample does not allow for further statistical analysis.

Distribution

Based on our phylogenetic analysis, this species is currently confirmed from three widely separated localities and elevations ranging from near sea level to > 1400 m: i) Kissama National Park, on the outskirts of Angola’s capital city, Luanda, in coastal western Angola (Luanda Province); ii) central Angola (Bié Province); and iii) the source of the Cuanavale, Cuito, Cuando and Quembo rivers (Moxico Province). The identity of other known Angolan localities for Breviceps (black diamonds) remain uncertain without additional sampling and genetic data (Fig. 1, Appendix 1; see Marques et al. 2018).

Genetic divergence

Breviceps ombelanonga differs from other species within the B. mossambicus group by net uncorrected mitochondrial p-distances of at least 9% (12S) and 11% (16S; Table 2), as well as unique nuclear haplotypes for both RAG1 and BDNF (Fig. 1B).

Habitat and natural history notes

The preferred habitat for B. ombelanonga ranges from typical western Angolan savannah, with sandy soils and vegetation dominated by Adansonia digitata, Euphorbia conspicua, Acacia welwitschii and Combretum sp., together with a good grass coverage (Grandvaux-Barbosa 1970), to dense Angolan wet miombo woodland in the east (Fig. 7). The type series was collected after gentle rains, either by hand or in traps. The holotype was first observed feeding on small, unidentified ants (family Formicidae). No information is available on egg deposit sites and clutch sizes. One of us (WC) has discovered remains of B. ombelanonga in the stomach contents of two snake species, Kladirostratus acutus (Psammophiidae; PEM R23450) and Causus bilineatus (Viperidae; PEM R23321) from the Cuando and Cuito River sources, respectively.

Figure 7. 

Photos of typical habitat of B. ombelanonga sp. nov.: A a view of the Kwanza River and bordering savannah, near the type locality, in Kissama National Park, Luanda Province B savannah near Embala Seque (14 km N of Cassumbi village), Bié Province C Cuanavale River source lake and associated miombo savannah woodland. Photographs by LMPC (A, B) and WC (C).

Etymology

The name ombelanonga is a derived combination of two words in Umbundu, a native Angolan language, for rain (ombela) and frog (anonga). The species epithet is used as an invariable noun in apposition to the generic name.

Conservation status

Given that it appears widely distributed, we suggest that B. ombelanonga be included in the IUCN category of Least Concern. The type locality lies within Kissama National Park, which grants some legal protection from major habitat degradation and loss, though the park has recently experienced significant wildfires. Additionally, the paratype localities in southeastern Angola (visited during field activities related to the National Geographic Okavango Wilderness Project 2017) are relatively pristine and ecologically intact miombo savannah that comprise an area recently proposed for formal protection.

Breviceps ombelanonga sp. nov. represents a phylogenetically distinct evolutionary lineage that is an Angolan endemic apparently geographically isolated from its closest congeners (Fig. 1). It forms a clade with morphologically similar members of the B. mossambicus group but can be differentiated from its sister taxon, B. poweri, phylogenetically, morphologically, and acoustically (Figs 1–3). Unlike B. poweri, B. ombelanonga lacks pale spots along the flanks, a pale patch above the vent, and a short, dark band below the nares (Fig. 4). We also recovered high intraspecific genetic diversity among populations of B. ombelanonga, which for the most part exceeds the reported interspecific distances for some recently described Breviceps species (Minter 2003; Minter et al. 2017; see Nielsen et al. 2018). Given the limited morphological variation within the novel taxon (and the B. mossambicus group more broadly), we elected to conservatively consider these three disjunct populations as one taxon.

Further work is required to confirm the distributional range of B. ombelanonga, as well as whether it overlaps in distribution with either its sister taxon, B. poweri, or the more distantly related B. adspersus. Both occur in neighboring countries, B. poweri to the east/northeast (Zambia, Democratic Republic of Congo) and the B. adspersus to the south/southeast (Namibia, Botswana), and both have been suggested to occur in Angola (Ruas 2002; Marques et al. 2018; Baptista et al. 2019; Channing and Rödel 2019; Fig. 1A). Due to the amount of morphological similarity found among most members of the B. mossambicus group, identifying museum specimens to species is difficult without having genetic data with which to assign populations. Therefore, we have elected to leave the historical specimens from Angola as unassigned (see Appendix 1). Revisiting historical collection localities, or in some cases attempting to acquire ‘historical’ DNA sequence data from museum specimens, carries high priority and should help to illuminate the composition and distribution of Angola’s resident Breviceps species.

We are not the first to recognize the lack of morphological variation within members of this anuran clade, which has led to historical taxonomic confusion and invoking hybridization for specimens that failed to conform to often scant descriptions of the type specimens (Poynton 1964, 1982; Poynton and Broadley 1985; Minter et al. 2017). The only comprehensive molecular phylogenetic study to date failed to find support for hybridization (Nielsen et al. 2018). Furthermore, many recent studies have shown that species discovery is still ongoing within this group (Minter et al. 2017), and that species thought to be widespread are often species-complexes composed of taxa with much narrower geographic ranges (Nielsen et al. 2018). Future, fine-scale fieldwork efforts targeting the many undersampled regions across the subcontinent where the B. mossambicus species group is likely to occur, combined with population genetic/phylogenomic methods, will be necessary to better investigate the presence of hybridization within Breviceps. We are optimistic that future studies scrutinizing morphological data (both morphometric and anatomical, i.e., via CT-scanning) of large numbers of genotyped B. mossambicus group samples will reveal diagnostic morphological differences between species and/or populations that are otherwise difficult to discern by individual specimens (Fig. 2).

As mentioned above, there is considerable genetic structure within B. ombelanonga, as well as among the four most closely related members of the B. mossambicus group (Fig. 1B, C). The Great Escarpment is a major topographical feature of southern Africa that separates the central plateau from coastal plains semi-continuously from Angola in the northwest, south through Namibia and South Africa, before petering out along the border of Zimbabwe and Mozambique in the northeast. This feature is coincident with changes in habitat and climate as one moves from the coast inland, and is consequently reflected in the distribution and diversification of various organisms (Clark et al. 2011; Nielsen et al. 2018). The western and central populations of B. ombelanonga, for example, are separated by the escarpment, although further study is needed to verify that the genetic structure we observed (between all three populations) is not just an effect of isolation by distance, compounded by limited sampling. Unfortunately, this is not unique to the B. mossambicus group. Many recent studies on other herpetofauna have stated that large sampling gaps across sub-Saharan Africa may cause misleading biogeographic conclusions (Medina et al. 2016; Jongsma et al. 2018). The central and eastern localities of B. ombelanonga, as well as the latter from either B. adspersus or B. poweri, may be separated by drainage basins; however, with no contemporary sampling across regions spanning hundreds of kilometers, it is difficult to test these broad biogeographic hypotheses. Many recent initiatives have improved the current state of knowledge of Angola’s herpetofauna, as well as to identify priority areas for future field survey work (Ceríaco et al. 2014, 2016, 2018; Conradie et al. 2016; Heinicke et al. 2017; Marques et al. 2018; Baptista et al. 2019; Butler et al. 2019; Ernst et al. 2020), yet these efforts have still only scratched the surface. Additional, comprehensive field surveys, particularly those with focused/specialized efforts to record hard-to-find, seasonal, and/or fossorial taxa (e.g., by deploying pitfall traps, drift fence arrays, artificial refuges, etc., for an extended period of time or repeatedly throughout the year), should be priorities in the near future.

We are grateful to Mariana P. Marques, Adam Ferguson, Ben Marks, John Cavagnaro, Philip Pastor, Suzana Bandeira, Ilola Jorge, Alvaro “Varito” Baptista, Ninda Baptista, Kerllen Costa, James Harvey, Roger Bills, Götz Neef, and Luke Verburgt for their invaluable field assistance, support, advice, and companionship. John Cavagnaro provided the life photo of the holotype. Mohamad Beidoun assisted with molecular work. This work was supported in part by the US National Science Foundation (DEB-1556255, 1556559 and DEB-1556585 to DCB, AMB, and MPH), the JRS Biodiversity Foundation (to DCB and AMB), and the National Geographic Society (Okavango Wilderness Project EC0715-15 to WC). The funders had no role in study design, data collection, and analysis, the decision to publish, or manuscript preparation. Field components of this work were facilitated by a Memorandum of Understanding between the University of Florida and The National Institute for Biodiversity and Protected Areas (INBAC) in Angola. Material used in this study was exported under the following permit numbers: 083/INBAC.MINAMB/2016 (to Villanova University) and 31/GGPCC/2016 (to the PEM). Some of the molecular work was completed at Marquette University and SVN would like to thank the good Dr. Gamble for his generosity. We would lastly thank the reviewers, Drs. Louis Du Preez and Alan Channing, for their kind words and edits that helped improve this manuscript, as well as our subject editor, Dr. Angelica Crottini.

Additional Breviceps material examined

Breviceps sp.: Angola: Lunda Sul Province: Alto Chicapa (MD 5426, 5865); Rives du lac Calundo (MD 5599); Moxico Province: Cazombo (MD 5770; MCZ A-35892–893); Luso: Calombe (IICT 339–1959, 404–1959, 453-6–1959); Benguela Province: Benguela (BMNH 1906.10.8.10-11); Ebanga (MHNC 90.008, 90.009); Chimbassi [= Chimbasse] (ZSM 173/1953); Quissange (BMNH 1887.3.23.5); Huambo Province: Bimbi (MCZ A-23721).

B. adspersus : Botswana: Serowe (PEM A4800); Namibia: Damaraland (ZMB 6294 [lectotype]), Okahandja (PEM A4723); South Africa: Limpopo Province: Waterpoort (PEM A14226); Mpumalanga Province: Botshabelo (ZMB 10087 [paralectotype]); Northern Cape Province: Rooipoort (PEM A8001–2, PEM A9431, PEM A9433–4), Tswalu (PEM A9444), Kuruman River Reserve (PEM A13883).

B. mossambicus : Malawi: Mount Mulanje (PEM A7861); Mozambique: Cabo Delgado Province: Balama (PEM A11021); Nampula Province: Insula Mossambique (ZMB 75399–400 [syntypes]), Mount Namuli (PEM A11310), Mount Ribaue (PEM A11362), Ribaue town (PEM A13952, PEM A13956), Nagonha Village (PEM A6717); Niassa Province: Lichinga (PEM A14008); Zambezi Province: Mount Lico (PEM A13725–6); Tanzania (ZMB 24793).

B. poweri : Democratic Republic of the Congo: Lualaba Province: Kalakundi (PEM A8453–6); Haut-Katanga Province: Sakania (UF Herp 27586); Mozambique: Nampula Province: Ribaue town (PEM A13957); Zambia: Northern Province: Mporokoso (PEM A2794); Northwestern Province: Solwezi (CAS 196527); Zimbabwe: Melsetter (PEM A4735).